In a recent CNN feature, César de la Fuente, Presidential Assistant Professor in Bioengineering, Psychiatry, Microbiology, and in Chemical and Biomolecular Engineering commented on a study about a new type of antibiotic that was discovered with artificial intelligence:
“I think AI, as we’ve seen, can be applied successfully in many domains, and I think drug discovery is sort of the next frontier.”
The de la Fuente lab uses machine learning and biology to help prevent, detect, and treat infectious diseases, and is pioneering the research and discovery of new antibiotics.
Myeloma cells producing monoclonal proteins of varying types, created by Scientific Animations under the Creative Commons Attributions-Share Alike International 4.0 License
Multiple myeloma is an incurable bone marrow cancer that kills over 100,000 people every year. Known for its quick and deadly spread, this disease is one of the most challenging to address. As these cancer cells move through different parts of the body, they mutate, outpacing possible treatments. People diagnosed with severe multiple myeloma that is resistant to chemotherapy typically survive for only three to six months. Innovative therapies are desperately needed to prevent the spread of this disease and provide a fighting chance for those who suffer from it.
Michael Mitchell, J. Peter and Geri Skirkanich Assistant Professor of Innovation in Bioengineering (BE), and Christian Figueroa-Espada, doctoral student in BE at the University of Pennsylvania School of Engineering and Applied Science, created an RNA nanoparticle therapy that makes it impossible for multiple myeloma to move and mutate. The treatment, described in their study published in PNAS, turns off a cancer-attracting function in blood vessels, disabling the pathways through which multiple myeloma cells travel.
By shutting down this “chemical GPS” that induces the migration of cancer cells, the team’s therapy stops the spread of multiple myeloma, helping to eliminate it altogether.
“Macrophages killing cancer cell” photographed by Susan Arnold.
By silencing the molecular pathway that prevents macrophages from attacking our own cells, Penn Engineers have manipulated these white blood cells to eliminate solid tumors.
Cancer remains one of the leading causes of death in the U.S. at over 600,000 deaths per year. Cancers that form solid tumors such as in the breast, brain or skin are particularly hard to treat. Surgery is typically the first line of defense for patients fighting solid tumors. But surgery may not remove all cancerous cells, and leftover cells can mutate and spread throughout the body. A more targeted and wholistic treatment could replace the blunt approach of surgery with one that eliminates cancer from the inside using our own cells.
Dennis Discher, Robert D. Bent Professor in Chemical and Biomolecular Engineering, Bioengineering, and Mechanical Engineering and Applied Mechanics, and postdoctoral fellow, Larry Dooling, provide a new approach in targeted therapies for solid tumor cancers in their study, published in Nature Biomedical Engineering. Their therapy not only eliminates cancerous cells, but teaches the immune system to recognize and kill them in the future.
“Due to a solid tumor’s physical properties, it is challenging to design molecules that can enter these masses,” says Discher. “Instead of creating a new molecule to do the job, we propose using cells that ‘eat’ invaders – macrophages.”
Macrophages, a type of white blood cell, immediately engulf and destroy – phagocytize – invaders such as bacteria, viruses, and even implants to remove them from the body. A macrophage’s innate immune response teaches our bodies to remember and attack invading cells in the future. This learned immunity is essential to creating a kind of cancer vaccine.
But, a macrophage can’t attack what it can’t see.
“Macrophages recognize cancer cells as part of the body, not invaders,” says Dooling. “To allow these white blood cells to see and attack cancer cells, we had to investigate the molecular pathway that controls cell-to-cell communication. Turning off this pathway – a checkpoint interaction between a protein called SIRPa on the macrophage and the CD47 protein found on all ‘self’ cells – was the key to creating this therapy.”
Multiple members in the biophysical engineering lab lead by Dennis Discher, including co-lead author and postdoctoral fellow and Penn Bioengineering alumnus Jason Andrechak and Bioengineering Ph.D. student Brandon Hayes, contributed to this study. The research was funded by grants from the National Heart, Lung, and Blood Institute and the National Cancer Institute, including the Physical Sciences Oncology Network, of the US National Institutes of Health.
Neil Sheppard, Adjunct Associate Professor of Pathology and Laboratory Medicine in the Perelman School of Medicine, and David Mai, a Bioengineering graduate student in the School of Engineering and Applied Science, explained the findings of their recent study, which offered a potential strategy to improve T cell therapy in solid tumors, to the European biotech news website Labiotech.
Machine learning (ML) programs computers to learn the way we do – through the continual assessment of data and identification of patterns based on past outcomes. ML can quickly pick out trends in big datasets, operate with little to no human interaction and improve its predictions over time. Due to these abilities, it is rapidly finding its way into medical research.
People with breast cancer may soon be diagnosed through ML faster than through a biopsy. Those suffering from depression might be able to predict mood changes through smart phone recordings of daily activities such as the time they wake up and amount of time they spend exercising. ML may also help paralyzed people regain autonomy using prosthetics controlled by patterns identified in brain scan data. ML research promises these and many other possibilities to help people lead healthier lives.
But while the number of ML studies grow, the actual use of it in doctors’ offices has not expanded much past simple functions such as converting voice to text for notetaking.
The limitations lie in medical research’s small sample sizes and unique datasets. This small data makes it hard for machines to identify meaningful patterns. The more data, the more accuracy in ML diagnoses and predictions. For many diagnostic uses, massive numbers of subjects in the thousands would be needed, but most studies use smaller numbers in the dozens of subjects.
But there are ways to find significant results from small datasets if you know how to manipulate the numbers. Running statistical tests over and over again with different subsets of your data can indicate significance in a dataset that in reality may be just random outliers.
This tactic, known as P-hacking or feature hacking in ML, leads to the creation of predictive models that are too limited to be useful in the real world. What looks good on paper doesn’t translate to a doctor’s ability to diagnose or treat us.
These statistical mistakes, oftentimes done unknowingly, can lead to dangerous conclusions.
To help scientists avoid these mistakes and push ML applications forward, Konrad Kording, Nathan Francis Mossell University Professor with appointments in the Departments of Bioengineering and Computer and Information Science in Penn Engineering and the Department of Neuroscience at Penn’s Perelman School of Medicine, is leading an aspect of a large, NIH-funded program known as CENTER – Creating an Educational Nexus for Training in Experimental Rigor. Kording will lead Penn’s cohort by creating the Community for Rigor which will provide open-access resources on conducting sound science. Members of this inclusive scientific community will be able to engage with ML simulations and discussion-based courses.
“The reason for the lack of ML in real-world scenarios is due to statistical misuse rather than the limitations of the tool itself,” says Kording. “If a study publishes a claim that seems too good to be true, it usually is, and many times we can track that back to their use of statistics.”
Such studies that make their way into peer-reviewed journals contribute to misinformation and mistrust in science and are more common than one might expect.
Brian Litt, Professor in Bioengineering in Penn Engineering and in Neurology in the Perelman School of Medicine, spoke to Neurology Today about the advances in technology for detecting and forecasting seizures.
The Litt Lab for Translational Neuroengineering translates neuroengineering research directly into patient care, focusing on epilepsy and a variety of research initiatives and clinical applications.
“Dr. Litt’s group is working with one of a number of startups developing ‘dry’ electrode headsets for home EEG monitoring. ‘They are still experimental, but they’re getting better, and I’m really optimistic about the possibilities there.'”
Brain development does not occur uniformly across the brain, but follows a newly identified developmental sequence, according to a new Penn Medicine study. Brain regions that support cognitive, social, and emotional functions appear to remain malleable—or capable of changing, adapting, and remodeling—longer than other brain regions, rendering youth sensitive to socioeconomic environments through adolescence. The findings are published in Nature Neuroscience.
Researchers charted how developmental processes unfold across the human brain from the ages of 8 to 23 years old through magnetic resonance imaging (MRI). The findings indicate a new approach to understanding the order in which individual brain regions show reductions in plasticity during development.
Brain plasticity refers to the capacity for neural circuits—connections and pathways in the brain for thought, emotion, and movement—to change or reorganize in response to internal biological signals or the external environment. While it is generally understood that children have higher brain plasticity than adults, this study provides new insights into where and when reductions in plasticity occur in the brain throughout childhood and adolescence.
The findings reveal that reductions in brain plasticity occur earliest in “sensory-motor” regions, such as visual and auditory regions, and occur later in “associative” regions, such as those involved in higher-order thinking (problem solving and social learning). As a result, brain regions that support executive, social, and emotional functions appear to be particularly malleable and responsive to the environment during early adolescence, as plasticity occurs later in development.
“Studying brain development in the living human brain is challenging. A lot of neuroscientists’ understanding about brain plasticity during development actually comes from studies conducted with rodents. But rodent brains do not have many of what we refer to as the association regions of the human brain, so we know less about how these important areas develop,” says corresponding author Theodore D. Satterthwaite,the McLure Associate Professor of Psychiatry in the Perelman School of Medicine, and director of the Penn Lifespan Informatics and Neuroimaging Center (PennLINC).
Throughout his career, Brian Litt has fabricated tools that support international collaboration, produced findings that have led to significant breakthroughs, and mentored the next generation of researchers tackling neurological disorders. (Image: iStock Photo/Alona Siniehina)
When Brian Litt of the Perelman School of Medicine and School of Engineering and Applied Science began treating patients as a neurologist, he found that the therapies and treatments for epilepsy were mostly reliant on traditional pharmacological interventions, which had limited success in changing the course of the disease.
People with epilepsy are often prescribed anti-seizure medications, and, while they are effective for many, about 30% of patients still continue to experience seizures. Litt sought new ways to offer patients better treatment options by investigating a class of devices that electronically stimulate cells in the brain to modulate activity known as neurostimulation devices.
Litt’s research on implantable neurostimulation devices has led to significant breakthroughs in the technology and has broadened scientists’ understanding of the brain. This work started not long after he came to Penn in 2002 with licensing algorithms to help drive a groundbreaking device by NeuroPace, the first closed-loop, responsive neurostimulator to treat epilepsy.
Building on this work, Litt noted in 2011 how the implantable neurostimulation devices being used at the time had rigid wires that didn’t conform to the brain’s surface, and he received support from CURE Epilepsy to accelerate the development of newer, flexible wires to monitor and stimulate the brain.
“CURE is one of the epilepsy community’s most influential funding organizations,” Litt says. “Their support for my lab has been incredibly helpful in enabling the cutting-edge research that we hope will change epilepsy care for our patients.”
A new Penn Medicine preclinical study demonstrates a simultaneous ‘knockout’ of two inflammatory regulators boosts T cell expansion to attack solid tumors.
by Meagan Raeke
Image: Courtesy of Penn Medicine News
A new approach that delivers a “one-two punch” to help T cells attack solid tumors is the focus of a preclinical study by researchers from the Perelman School of Medicine. The findings, published in the Proceedings of the National Academy of Sciences, show that targeting two regulators that control gene functions related to inflammation led to at least 10 times greater T cell expansion in models, resulting in increased anti-tumor immune activity and durability.
“We want to unlock CAR T cell therapy for patients with solid tumors, which include the most commonly diagnosed cancer types,” says June, the new study’s senior author. “Our study shows that immune inflammatory regulator targeting is worth additional investigation to enhance T cell potency.”
One of the challenges for CAR T cell therapy in solid tumors is a phenomenon known as T cell exhaustion, where the persistent antigen exposure from the solid mass of tumor cells wears out the T cells to the point that they aren’t able to mount an anti-tumor response. Engineering already exhausted T cells from patients for CAR T cell therapy results in a less effective product because the T cells don’t multiply enough or remember their task as well.
Previous observational studies hinted at the inflammatory regulator Regnase-1 as a potential target to indirectly overcome the effects of T cell exhaustion because it can cause hyperinflammation when disrupted in T cells—reviving them to produce an anti-tumor response. The research team, including lead author David Mai, a bioengineering graduate student in the School of Engineering and Applied Science, and co-corresponding author Neil Sheppard, head of the CCI T Cell Engineering Lab, hypothesized that targeting the related, but independent Roquin-1 regulator at the same time could boost responses further.
“Each of these two regulatory genes has been implicated in restricting T cell inflammatory responses, but we found that disrupting them together produced much greater anti-cancer effects than disrupting them individually,” Mai says. “By building on previous research, we are starting to get closer to strategies that seem to be promising in the solid tumor context.”
With OCTOPUS, Dan Huh’s team has significantly advanced the frontiers of organoid research, providing a platform superior to conventional gel droplets. OCTOPUS splits the soft hydrogel culture material into a tentacled geometry. The thin, radial culture chambers sit on a circular disk the size of a U.S. quarter, allowing organoids to advance to an unprecedented degree of maturity.
When it comes to human bodies, there is no such thing as typical. Variation is the rule. In recent years, the biological sciences have increased their focus on exploring the poignant lack of norms between individuals, and medical and pharmaceutical researchers are asking questions about translating insights concerning biological variation into more precise and compassionate care.
What if therapies could be tailored to each patient? What would happen if we could predict an individual body’s response to a drug before trial-and-error treatment? Is it possible to understand the way a person’s disease begins and develops so we can know exactly how to cure it?
Dan Huh, Associate Professor in the Department of Bioengineering at the University of Pennsylvania’s School of Engineering and Applied Science, seeks answers to these questions by replicating biological systems outside of the body. These external copies of internal systems promise to boost drug efficacy while providing new levels of knowledge about patient health.
An innovator of organ-on-a-chip technology, or miniature copies of bodily systems stored in plastic devices no larger than a thumb drive, Huh has broadened his attention to engineering mini-organs in a dish using a patient’s own cells.
Machine learning (ML) programs computers to learn the way we do – through the continual assessment of data and identification of patterns based on past outcomes. ML can quickly pick out trends in big datasets, operate with little to no human interaction and improve its predictions over time. Due to these abilities, it is rapidly finding its way into medical research.
